ud_srl-enpt_xlmr-large

XLM-R large fine-tuned in Portuguese Universal Dependencies and English and Portuguese semantic role labeling

Model description

This model is the xlm-roberta-large fine-tuned first on the Universal Dependencies Portuguese dataset, then fine-tuned on the CoNLL formatted OntoNotes v5.0 and then fine-tuned on the PropBank.Br data. This is part of a project from which resulted the following models:

• liaad/srl-pt_bertimbau-base
• liaad/srl-pt_bertimbau-large
• liaad/srl-pt_xlmr-base
• liaad/srl-pt_xlmr-large
• liaad/srl-pt_mbert-base
• liaad/srl-en_xlmr-base
• liaad/srl-en_xlmr-large
• liaad/srl-en_mbert-base
• liaad/srl-enpt_xlmr-base
• liaad/srl-enpt_xlmr-large
• liaad/srl-enpt_mbert-base
• liaad/ud_srl-pt_bertimbau-large
• liaad/ud_srl-pt_xlmr-large
• liaad/ud_srl-enpt_xlmr-large

For more information, please see the accompanying article (See BibTeX entry and citation info below) and the project's github.

Intended uses & limitations

How to use

To use the transformers portion of this model:

from transformers import AutoTokenizer, AutoModel

tokenizer = AutoTokenizer.from_pretrained("liaad/ud_srl-enpt_xlmr-large")
model = AutoModel.from_pretrained("liaad/ud_srl-enpt_xlmr-large")

To use the full SRL model (transformers portion + a decoding layer), refer to the project's github.

Limitations and bias

• This model does not include a Tensorflow version. This is because the "type_vocab_size" in this model was changed (from 1 to 2) and, therefore, it cannot be easily converted to Tensorflow.
• The model was trained only for 10 epochs in the Universal Dependencies dataset.
• The model was trained only for 5 epochs in the CoNLL formatted OntoNotes v5.0.
• The English data was preprocessed to match the Portuguese data, so there are some differences in role attributions and some roles were removed from the data.

Training procedure

The model was trained on the Universal Dependencies Portuguese dataset; then on the CoNLL formatted OntoNotes v5.0; then on Portuguese semantic role labeling data (PropBank.Br) using 10-fold Cross-Validation. The 10 resulting models were tested on the folds as well as on a smaller opinion dataset "Buscapé". For more information, please see the accompanying article (See BibTeX entry and citation info below) and the project's github.

Eval results

BibTeX entry and citation info

@misc{oliveira2021transformers,
      title={Transformers and Transfer Learning for Improving Portuguese Semantic Role Labeling}, 
      author={Sofia Oliveira and Daniel Loureiro and Alípio Jorge},
      year={2021},
      eprint={2101.01213},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}